| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Diagnostic Radiology, Yale University School of Medicine, 333
Cedar St., New Haven, CT 06520.
2
Present address: Department of Radiology, New York Presbyterian Hospital, 525
E. 68th St., New York, NY 10021.
Received August 3, 2000;
accepted after revision October 24, 2000.
Presented at the annual meeting of the Radiological Society of North
America, Chicago, November 1999.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Reports from MR examinations of 103 patients who underwent MR imaging of the breast to evaluate questionable mammographic findings were retrospectively reviewed. We identified cases that had focal enhancing lesions without a corresponding mammographic or palpable abnormality. Clinical history, mammograms, MR images, and follow-up information were reviewed.
RESULTS. Of the 103 patients, 30 (29%) had incidental foci of enhancement. These women were significantly younger, more often premenopausal, and more likely to have dense breasts than those who did not have incidental foci. Tissue confirmation of the incidental foci was available for seven patients, mammographic follow-up was available for a mean interval of 22 months for 22 patients, and no follow-up was available for one. Cancer at the incidental sites was diagnosed in one of the 30 patients with multiple foci. She was also shown to have cancer at the site originally questioned mammographically (index site). None of the remaining patients has had a diagnosis of malignancy at the incidental sites.
CONCLUSION. Incidental enhancing foci are common in women undergoing breast MR imaging for questionable findings on mammography, occurring in 29% of our patients. Our results suggest that unless malignancy is diagnosed elsewhere in the breast, these incidental foci are unlikely to be malignant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A large portion of the breast MR examinations at our institution are performed as a problem-solving step for patients who have an uncertain finding on mammography. Often, these are high-risk patients in whom a more immediate answer is desired, rather than the delayed answer provided by short-interval mammographic follow-up. The purpose of this study was to determine the frequency and significance of these incidental foci of enhancement on breast MR images that are performed to evaluate equivocal mammographic findings.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The reasons for performing MR imaging were as follows: In 36 patients, a focal area of asymmetric density was seen on mammography, but its significance was unclear. These densities were usually new or increased from the previous mammograms, often in the setting of newly instituted hormone replacement therapy, but were otherwise relatively benignappearing. In 29 patients, a vague area of architectural distortion or increased density was questioned but could not be definitively confirmed with additional images. In 26 patients, a change was questioned at the site of a previous lumpectomy or biopsy with benign results. In the final 12 patients, a more suspicious abnormality was seen but could not be targeted sufficiently to allow biopsy to be performed. Although somewhat suspicious, the findings in these patients were vague and were considered equivocal. The mammographic abnormality that prompted the MR examination is referred to as the index lesion.
All imaging was performed using a 1.5-T MR scanner (Signa; General Electric Medical Systems, Milwaukee, WI) and a commercially available receiveonly phased array breast coil (Medrad, Pittsburgh, PA). Patients were imaged in the prone position with both breasts hanging dependently. Compression was not used. Gadolinium-enhanced images were obtained using a dose of 0.1 mmol/kg of gadopentetate dimeglumine (Magnevist, Berlex Laboratories, Richmond, CA; or Gadodiamide, Nycomed Amersham, Princeton, NJ). IV lines were placed in the antecubital fossa before the initiation of imaging. A sufficient length of IV tubing was used so that the patient could remain positioned in the scanner during contrast medium injection.
The following sequences were obtained for all patients: axial fast spoiled gradient-echo (SPGR) to localizer sequence (TR/TE, 200/4.2; flip angle, 75°; bandwidth, 32 kHz; matrix, 256 x 160; 1 signal average; field of view, 30 cm; section thickness, 5 mm; intersection spacing, 2 mm), followed by sagittal conventional spin-echo T1-weighted images (TR/TE, 600/minimum; matrix, 256 x 192; 1 signal average; field of view, 14-18 cm; 3-mm section thickness, interleaved) and sagittal fast short tau inversion recovery (STIR) images (TR/TE, 4000/100; inversion time, 130 msec; echo train length, 8; bandwidth, 32 kHz; matrix, 256 x 192; 2 signal averages; field of view, 14-18 cm; 5-mm section thickness; 1-mm intersection spacing). Unenhanced and gadolinium-enhanced images were then obtained using a chemical selective fat-suppressed three-dimensional fast SPGR sequence (TR/TE, minimum/minimum; 12 or 28 sections depending on breast size and region of interest; flip angle, 45°; matrix, 256 x 224; 2 signal averages; field of view, 14-18 cm; section thickness, 2-4 mm). Scanning time for the latter sequence was approximately 5 min. The injection of gadolinium was timed so that contrast material reached the breast at or slightly before the center of the data set. When images are obtained in this manner, the contrast characteristics of the images are similar to those obtained in the early phase of a dynamically enhanced scan. This is similar in theory to the timing of the contrast injection for MR angiography of the aorta, in which the contrast material reaches the aorta at the center of the data set. In most patients, injection was begun approximately 1 min 45 sec into the data acquisition. A delayed scan was then immediately obtained with the same three-dimensional fast SPGR sequence.
An MR imaging—guided biopsy of suspicious lesions not visible or accessible by other techniques was performed using the freehand technique described by Brenner et al. [10].
Reports from MR imaging of the 103 patients who constituted our study population were retrospectively reviewed by two of the authors to identify cases in which at least one mammographically occult focal enhancing lesion was seen. At our institution, all interpreters routinely include in their reports whether a lesion on MR imaging corresponds to an abnormality seen on mammography. An attending radiologist experienced in MR imaging and another experienced in mammography collaboratively reviewed the MR images with the mammograms to retrospectively confirm that the lesion seen on MR imaging was indeed mammographically occult. Any discrete focus of enhancement on the delayed images that was not thought to be normal breast parenchyma was considered positive for an abnormality. We included all enhancing lesions that did not have a mammographic correlate even in retrospect. The location, number, and size of the incidental enhancing lesions were recorded. Enhancement of the lesions was characterized as homogeneous, heterogeneous, rim-enhancing, or having dark internal septations.
Treatment of the incidental lesions was decided jointly by the interpreting MR and breast imaging radiologists using several criteria: whether cancer was diagnosed at the index lesion, the patient's risk factors for breast cancer, the size and number of the incidental foci, and the morphology of the enhancing foci. In general, enhancing foci in women with cancer at the index site, solitary lesions, and foci with spiculated margins or rim enhancement were viewed with greater suspicion. Multiple, subcentimeter, homogeneously enhancing foci were usually managed with mammographic follow-up. For patients with multiple small foci and a single dominant focus, the dominant focus was generally considered separately as if it were a solitary lesion.
Clinical histories were reviewed, and note was made of any personal history of breast cancer or its precursors, a significant family history of breast cancer, the patient's age and menopausal status, and any hormone replacement therapy. Reports from sonographic examinations, follow-up mammograms, and MR examinations, and the pathology reports for patients who underwent tissue sampling, were reviewed. Whenever possible, the pathologic specimens or images from biopsy procedures were reviewed to determine as closely as possible the location of the MR imaging abnormalities relative to the biopsied or resected tissue.
Contingency tables were analyzed using the chi-square test with Yates correction and the Student's t test. Statistical significance was accepted if the p value was equal to or less than 0.05. Statistical values were calculated using SYSTAT statistical software, version 5.2 (Systat, Evanston, IL).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
In 12 of 30 patients, the incidental foci were multiple (>3) and ranged in size from 1 to 5 mm (Fig. 1). Seven patients had a single incidental focus of enhancement measuring 2-11 mm. In five patients, multiple small areas of enhancement (1-5 mm) were seen, with a single dominant focus of enhancement ranging from 6 to 14 mm. Two foci, ranging from 2 to 8 mm, were found in five patients; and one patient had three foci, each measuring 8 mm (Table 3). The enhancement pattern was heterogeneous in three lesions, rim-enhancing in two, dark internal septations in one, and homogeneously enhancing in the remainder.
|
|
The patients were treated as follows: Sonography of the breast directed
toward the site of the incidental enhancement was performed after the MR
examination for six patients with lesions thought to be of sufficient size to
be evaluated with sonography (
6 mm). A corresponding solid mass was found
in two of these six patients. The mass had dark internal septations on MR
images and sonographic features typical of a fibroadenoma in one of these
patients, and management with follow-up rather than biopsy was elected.
Eighteen months after the MR examination, two new suspicious areas of
calcification developed on mammography at sites thought, after careful review
of the mammograms with the MR images, to be separate from the MR imaging
abnormality. These sites were biopsied and revealed invasive carcinoma. That
patient went on to have a mastectomy, and no additional sites of cancer were
found. Fibrocystic change was present, but the lesion causing the MR imaging
abnormality was not identified in the specimen. A sonographically guided
biopsy was performed for the second patient with a 10-mm solid mass
corresponding to a dominant focus of MR imaging enhancement; the biopsy
revealed a fibroadenoma. Additional 1 - to 4-mm incidental lesions were
followed up, with mammographic stability for 16 months. No focal abnormalities
were identified on sonography at the sites of the MR imaging abnormality in
the remaining four patients. Two of the patients with no mass seen on
sonography had a second MR examination, one showing no change in the lesion at
6 months, and the other showing resolution of the lesion at 2 months. All four
of these patients received subsequent mammographic follow-up with stable
findings after being monitored for 22-32 months (mean, 27 months).
MR imaging—guided biopsy of the incidental focus was recommended for four patients: two because cancer had been diagnosed at the index site, one because of a family history of breast cancer and a high level of patient anxiety, and one because of a family history of breast cancer and irregular margins of the solitary focus. In two of these four patients—one with cancer at the index site and one without—MR imaging—guided biopsy was performed with negative results. In the other two patients—one with cancer at the index site and one without—the incidental enhancing focus was no longer seen on the second MR examination performed at the time of the planned biopsy.
Four patients with a diagnosis of cancer at the index site and the one patient previously discussed who developed cancer at a separate site during the follow-up period underwent mastectomy. One patient who ultimately underwent mastectomy for cancer diagnosed at the index site initially had an MR imaging—guided biopsy with benign results of a single incidental enhancing lesion, discussed previously. The site of the biopsy was identified in the mastectomy specimen, and it was confirmed that the area was free of cancer. For two patients with cancer diagnosed at the index site, the quadrant containing the incidental foci was free of tumor at mastectomy. The fourth patient had two 8-mm incidental foci, one that enhanced homogeneously and one with rim-enhancement. Extensive ductal carcinoma in situ (DCIS) was found throughout the breast, including the areas of the two incidental foci.
Therefore, a total of seven patients had tissue confirmation of the incidental foci, one diagnosed with biopsy under sonographic guidance, one with biopsy under MR imaging guidance, one with biopsy under MR imaging guidance followed by mastectomy, and four at mastectomy. Only one of the seven, a patient with two incidental foci and with cancer diagnosed at the index site, had malignancy at the incidental sites.
Of the 23 patients who did not have tissue confirmation, 22 had subsequent mammographic examination, including seven patients who were also examined with other modalities. None developed a mammographic abnormality to correspond with the incidental enhancing foci found on MR imaging examination during follow-up ranging from 2 to 42 months (mean, 22 months). This group included one patient with a rim-enhancing incidental focus, whose mammograms were stable for 25 months. No follow-up was available for one patient.
In addition to mammographic follow-up, MR imaging follow-up was performed for six women. The incidental lesions showed stability in two patients who were followed up for 6-18 months. The foci disappeared at follow-up MR imaging in three patients (one at 2 months, one at 6 months, and one at 12 months) (Fig. 2A,2B). The sixth patient had three incidental sites, one of which was biopsied under MR imaging guidance and showed benign tissue (discussed previously), and the other two lesions were no longer seen on the second MR examination performed 1 month later at the time of the biopsy.
|
|
In summary, of 103 patients undergoing breast MR imaging for further evaluation of equivocal mammographic findings (index lesions), 30 patients had incidental foci of enhancement at other sites in the breast. In only a single patient were the incidental foci of enhancement shown to contain cancer, and that occurred in a patient with DCIS at the index lesion and, in fact, throughout the remainder of her breast.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It is also not known what role a patient's age plays in the prevalence of enhancing foci in an actual patient population. We evaluated the breast MR images of patients undergoing examination for equivocal findings on mammography to determine how often incidental enhancing foci occur in this patient population, which has become a major subset of the patients referred for breast MR imaging at our institution. We showed that for women without known malignancy who are under-going breast MR examination for questionable findings on mammography, incidental enhancing foci are common, occurring in 29% of patients.
Factors associated with a greater likelihood of finding incidental enhancing foci included younger age (p < 0.01) and being premenopausal (p < 0.02). The distribution of breast parenchymal patterns was also different between our groups of patients (p < 0.01), with (in general) patients with denser breasts having a greater likelihood of incidental foci. We found no significant association between incidental foci and a personal history of breast cancer (p < 0.09), a family history of breast cancer (p < 0.38), or hormone replacement therapy (p < 0.26).
The significance of incidental enhancing lesions may differ on the basis of the population of patients imaged. Lawrence et al. [12] proposed a statistical model for the interpretation of incidental lesions for women who are "typical members of the population." Using metaanalysis of the literature on the sensitivity and specificity of MR imaging, mammography, and clinical breast examination, and calculating the United States population prevalence of breast cancer from registry data, these researchers concluded that negative findings on mammography and on clinical breast examination lower the probability of disease more than positive findings on MR imaging raise the probability. Until now, actual patients have not been studied. If we consider only the subset of the patients in our study who conform to the population analyzed by Lawrence et al., our results concur with their conclusions. None of our patients with negative findings for index lesions have been shown, to date, to have cancer in any of the enhancing foci found on MR imaging that were occult to other modalities.
But our patient population was selected and included many women who are at high risk for cancer and who would be excluded from the analysis performed by Lawrence et al. [12]. Our study population included women in whom cancer was diagnosed at the index lesion (n = 5); those who had cancer diagnosed at other sites in the breast during follow-up (n = 1); and those with a personal history of invasive breast cancer (n = 9), lobular carcinoma in situ (n = 2), or atypical lobular hyperplasia (n = 1). Negative findings on MR imaging in these high-risk patients allow more confident management on short-interval follow-up rather than on biopsy [1]. But, just as relatively benign-appearing findings on mammography may be viewed with more suspicion in high-risk patients, so are the enhancing foci on MR imaging. Enhancing lesions in these patients are not easily dismissed on statistical grounds alone.
Even in our high-risk patients, only one patient has proven so far to have cancer at a site of incidental enhancement. This patient was premenopausal and had a history of biopsy-proven lobular carcinoma in situ. Her MR imaging examination showed abnormal enhancement at the index lesion and detected two incidental lesions of 8 mm each, one of which enhanced homogeneously and the other with rim enhancement. A biopsy revealed DCIS at the index site. At mastectomy, DCIS was found throughout the breast. After consultation with the interpreting pathologist and correlation with the images, it was concluded that the sites of incidental enhancement on the MR images were indeed involved with DCIS.
The significance of the number, size, shape, and enhancement characteristics of the lesions we detected has not been fully determined. For the purposes of our study, we included all incidental enhancing foci; we did not use morphologic features such as shape or border, or the timing or internal pattern of enhancement, to exclude lesions. The designation of a "lesion" is also somewhat subjective, because the differentiation of a focus of enhancement from normal enhancing parenchyma can be difficult. Decisions about the management of patients with incidental foci, however, were made in conjunction with information about the morphologic features and enhancement patterns of the lesions as well as the patient's risk factors and findings at the index site. Viehweg et al. [13] suggested in their interpretation guidelines that a strategy of highest possible sensitivity be used for lesions suspected on the basis of conventional imaging, but that the strategy should switch to one that maximizes specificity for lesions detected only with MR imaging. These researchers propose that a carcinoma misclassified as benign under this strategy could still be recognized by growth on subsequent MR imaging.
One limitation of our study involves the hormonal status of our patients. At our institution, the scheduling of breast MR imaging has not routinely been done according to the phase in the menstrual cycle of premenopausal women. A study of healthy premenopausal volunteers imaged during each week of the menstrual cycle revealed reversibility of 14 (61%) of the 23 focal enhancing areas detected on breast MR imaging, with a significant decrease in the conspicuity and prevalence of contrast-enhancing lesions, in week 2 of the menstrual cycle [11]. The three women in our study whose incidental lesions disappeared on follow-up MR imaging were premenopausal. Perhaps imaging in the second week or at least the middle of the menstrual cycle, as advocated by Viehweg et al. [13] and Muller-Schimpfle et al. [14], might reduce the number of cases with incidental abnormalities.
The role that hormone replacement therapy may have on incidental enhancement in breast MR imaging also has yet to be defined. We did not find an association, but the number of women on hormone replacement therapy in our study is fairly small; and because this was a retrospective study, the number of patients receiving hormone replacement and the type of hormone therapy were not controlled. With further study of postmenopausal women, it may be seen that hormone replacement therapy has an effect on incidental enhancement, as is seen with the menstrual cycle.
Another limitation is the relatively short follow-up interval. Most of our patients were followed up mammographically, with relatively few receiving MR imaging follow-up. Mammographic follow-up presupposes that any true malignancy would not remain mammographically occult over time. It is possible that incidental lesions will require longer mammographic follow-up to ensure benignity. Short-interval MR imaging follow-up for more patients could also prove beneficial to our under-standing of the natural history of these incidental lesions. Further study, including a cost-benefit analysis, is needed in this area.
On the basis of our experience, we recommend the following steps when MR imaging performed to evaluate one suspicious area shows an unexpected area of enhancement elsewhere. First, the mammograms should be collaboratively reviewed with radiologists experienced in MR imaging and in mammography to be certain that a corresponding lesion is not seen in retrospect on the mammogram. If necessary, additional mammography or directed sonography can be performed at the site, which may reveal characteristics that sway suspicion one way or the other. If a biopsy is determined to be necessary, it may be more easily performed under sonographic guidance.
For patients in whom sonography is not helpful, the decision regarding imaging follow-up or biopsy should be made after considering the morphologic and enhancement characteristics, the risk factors, and the anxiety level of the patient. Patients with more benign MR imaging characteristics can probably be safely followed up with mammography. We believe this is particularly true of patients with multiple enhancing foci of less than 1 cm. Repeating the MR examination in accordance with the menstrual cycle, if not already done, may spare some premenopausal patients an unnecessary biopsy [11]. If the index lesion findings prove to be negative, close follow-up with mammography or MR imaging may be reasonable if all other factors suggest benignity, because none of our patients with negative findings for index lesions were found to have cancer at the sites noted incidentally. However, for patients whose ultimate diagnosis is cancer in the same breast, the incidental lesions should be viewed with more suspicion.
In conclusion, incidental contrast-enhancing lesions on breast MR imaging performed for mammograms with equivocal findings are common and occur significantly more often in young women, premenopausal women, and women with dense breasts. Our data suggest that in the absence of cancer elsewhere in the breast, incidental foci of enhancement that do not have a mammographic, sonographic, or clinical correlate are unlikely to be malignant. If the morphology of the incidental focus is not suspicious, it seems reasonable to perform close follow-up of these patients with mammography and perhaps MR imaging, rather than immediate biopsy or excision.
Acknowledgments
We thank Robert C. Lange for providing assistance with statistical analysis
of the data.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  E. E. Deurloo, S. H. Muller, J. L. Peterse, A. P. E. Besnard, and K. G. A. Gilhuijs Clinically and Mammographically Occult Breast Lesions on MR Images: Potential Effect of Computerized Assessment on Clinical Reading Radiology, March 1, 2005; 234(3): 693 - 701. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. A. Berg, L. Gutierrez, M. S. NessAiver, W. B. Carter, M. Bhargavan, R. S. Lewis, and O. B. Ioffe Diagnostic Accuracy of Mammography, Clinical Examination, US, and MR Imaging in Preoperative Assessment of Breast Cancer Radiology, December 1, 2004; 233(3): 830 - 849. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. L. Rosen, K. L. Blackwell, J. A. Baker, M. S. Soo, R. C. Bentley, D. Yu, T. V. Samulski, and M. W. Dewhirst Accuracy of MRI in the Detection of Residual Breast Cancer After Neoadjuvant Chemotherapy Am. J. Roentgenol., November 1, 2003; 181(5): 1275 - 1282. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Teifke, H. A. Lehr, T. W. Vomweg, A. Hlawatsch, and M. Thelen Outcome Analysis and Rational Management of Enhancing Lesions Incidentally Detected on Contrast-Enhanced MRI of the Breast Am. J. Roentgenol., September 1, 2003; 181(3): 655 - 662. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. R. LaTrenta, J. H. Menell, E. A. Morris, A. F. Abramson, D. D. Dershaw, and L. Liberman Breast Lesions Detected with MR Imaging: Utility and Histopathologic Importance of Identification with US Radiology, June 1, 2003; 227(3): 856 - 861. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. G. A. Gilhuijs, E. E. Deurloo, S. H. Muller, J. L. Peterse, and L. J. Schultze Kool Breast MR Imaging in Women at Increased Lifetime Risk of Breast Cancer: Clinical System for Computerized Assessment of Breast Lesions—Initial Results Radiology, December 1, 2002; 225(3): 907 - 916. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
